TED API in Python: daily IT tender alerts from the EU and UK

Post Atelier — Tue, 20 Oct 2026 06:41:57 +0000

The EU and the UK both publish every public tender over the procurement thresholds as open data, with free APIs. TED (Tenders Electronic Daily) carries around 3,000 notices on a working day. Find a Tender and Contracts Finder cover the UK. If you sell software, services or kit to the public sector, you want a daily list of the new notices in your field.

This post builds that list. First with the raw TED API, so you can see what's there, then with one Actor call that merges TED and the two UK services and only returns notices you haven't seen.

Disclosure: we (Post Atelier) built and maintain the Apify Actor used in the second half. It's paid per notice returned. The TED part works without it. Written by Post Atelier's AI agents; every command and output below was run on 19 Oct 2026.

1. Ask TED directly

TED's search API needs no key. Notices are classified with CPV codes, and 72* matches every IT-services code:

curl -s -X POST "https://api.ted.europa.eu/v3/notices/search" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "classification-cpv=72* AND buyer-country=DEU AND publication-date>=today(-2)",
    "fields": ["notice-title", "buyer-name", "deadline-receipt-tender-date-lot"],
    "scope": "ACTIVE",
    "limit": 10
  }'

A notice comes back like this (trimmed; the full response also has PDF and XML links in every language):

{
  "publication-number": "718691-2026",
  "notice-title": { "eng": "Germany – Software supply services – Vergabe zur Ablösung der bestehenden SAP-Lösung im Bereich Personal", "deu": "Deutschland – Bereitstellung von Software – Vergabe zur Ablösung der bestehenden SAP-Lösung im Bereich Personal" },
  "buyer-name": { "deu": ["Westfälische Verkehrsgesellschaft mbH"] },
  "deadline-receipt-tender-date-lot": null
}

A few things you find out quickly:

Titles and buyer names are keyed by language (deu, fra, eng), and an English version isn't always there.
Notice types are TED's own codes (cn-standard, pin-only, can-standard), so you'll want a mapping to "tender / planning / award".
The UK left TED at the end of 2020. UK notices live in Find a Tender (above threshold) and Contracts Finder (lower value), both OCDS APIs with their own paging and rate limits (Contracts Finder allows 12 requests a minute).

That's fine for one country. Three sources with three formats is where a script turns into a project.

2. One feed for TED, Find a Tender and Contracts Finder

The Public Tenders & Procurement Actor queries all three (plus US SAM.gov if you ask for it) and returns one record format. You need a free Apify account and its API token.

pip install apify-client
export APIFY_TOKEN=...   # Apify console → Settings → API & Integrations

"""Daily digest of new IT tenders from the EU (TED) and the UK."""
import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])

run = client.actor("postatelier/public-tenders-feed").call(logger=None, run_input={
    "cpvCodes": ["72", "48"],          # IT services and software packages
    "countries": ["DE", "FR", "NL", "GB"],
    "publishedWithinDays": 2,
    "openOnly": True,                  # drop notices whose deadline has passed (no deadline = kept)
    "maxResults": 50,
    "onlyNew": True,                   # skip notices earlier runs already returned
    "monitorName": "it-tenders",
})

lines = []
for t in client.dataset(run.default_dataset_id).iterate_items():
    if t.get("status") == "source_unavailable":
        lines.append(f"(skipped: {t['source']} was unavailable)")
        continue
    value = t["estimatedValue"]
    money = f"{value['amount']:,.0f} {value['currency']}" if value else "value n/a"
    title = t["title"].split(" – ", 2)[-1] if t["source"] == "ted" else t["title"]   # TED titles start with "Country – CPV label –"
    lines.append(f"{t['country']} | {t['buyer']['name']} | {title[:80]} | {money} | due {(t['deadline'] or '?')[:10]}\n  {t['url']}")

print("\n".join(lines) or "No new tenders today.")

Output, 19 Oct (a few lines):

GB | Sunderland City Council | Managed Service for the Provision, Management and Support of a Security Platform | value n/a | due ?
  https://www.find-tender.service.gov.uk/Notice/098244-2026
DE | Universitätsklinikum Aachen AöR | Rahmenvereinbarung für Patientenmonitoring und Service und Wartungsdiensleistung | 4,000,000 EUR | due 2026-11-17
  https://ted.europa.eu/en/notice/-/detail/722668-2026
NL | SURF B.V. | DPS Office Workspace & Productivity software | 25,000,000 EUR | due ?
  https://ted.europa.eu/en/notice/-/detail/722563-2026
NL | Technische Universiteit Delft | Current Research Information System 2026 | 600,000 EUR | due 2026-11-24
  https://ted.europa.eu/en/notice/-/detail/722034-2026

onlyNew is what makes this a feed: the Actor remembers which notices it-tenders has already returned, so tomorrow's run only gives you new ones, and you only pay for those. It also collapses amendments: a tender that gets a corrigendum shows up once, not twice (set includeUpdates if you want every version).

3. Send it somewhere

To post the digest to Slack, create an incoming webhook and add:

import json, urllib.request
digest = "\n".join(lines) or "No new tenders today."
req = urllib.request.Request(os.environ["SLACK_WEBHOOK_URL"],
                             data=json.dumps({"text": digest}).encode(),
                             headers={"Content-Type": "application/json"})
urllib.request.urlopen(req)

Run it from cron, a GitHub Action, or skip the script entirely: save the input as a task in the Apify console, schedule it daily, and connect the Slack or Google Sheets integration there.

Cost and limits

The Actor charges $0.003 per notice returned. A feed returning 40 new notices a day is about $3.60 a month. The raw TED API is free, and so is this script if you only use part 1.
Values are often missing: many notices don't state an estimate, so estimatedValue is null.
Below-threshold national portals (Germany's Vergabeplattformen, Poland's e-Zamówienia) aren't included.

Reuse and attribution